The Functional Imaging Methods (FIM) section continues to develop cutting edge methods at the interface of technology development, processing advances, basic neurophysiology, and neuroscience applications. Specifically, this report describes progress in our research towards making MRI and fMRI more robust, more widely used, and more interpretable. Resting state oscillations: beat frequency mapping Carried out by Dan Handwerker. The methods for extracting resting state fluctuations are still evolving. The primary approaches include either using a seed voxel correlation approach very limited and labor intensive, yet very straightforward or using independent component analysis very efficient but not very clear as to what each component shows. We have come up with a novel method that is methodologically between these two approaches by making use of our observation that resting state connectivity varies over time. We are able to use an arbitrary reference function and observe how each voxel correlation with this varies. Each network has a specific frequency of variation. Separation of artifactual fluctuations from neuronally-relevant fluctuations Carried out primarily by Prantik Kundu, with the help of Souheil Inati. This is also in the context of collaboration with Ed Bullmore (co-advisor of Prantik). Independent component analysis (ICA) is commonly applied to fMRI time series to pull out patterns of activation. The primary problem with ICA is that the various components that come out are in random order and can vary considerably, so determining whats real or artifactual requires intensive user selection. We collect multi-echo EPI and then model the MR relaxation rate characteristics of the data to better separate real (i.e. related to changes in T2* and not another mechanism and coming from regions with T2* in the range of what gray matter should be) from artifactual components. Understanding the effect of the Valsalva maneuver on BOLD signal Carried out primarily by Paula Wu and Dan Handwerker. A common method for calibration of fMRI signal changes has been to induce a global change in flow, which then highlights variations in hemodynamics (namely baseline blood volume) which contribute to variations in BOLD contrast. We find that Valsalva enhances hypercapia effects and creates changes more rapidly. Future work involves exploring mechanisms behind these changes and applications of the Valsalva towards calibration. Comparison of resting-state connectivity assessment with DTI vs. fMRI Carried out by Sunil Narayan and Dan Handwerker. Diffusion tensor imaging tractography is a method for estimating white matter bundles in the human brain that produce a visual representation for the structural connectivity of different functional regions. Combining this technique with methods for assessing functional connectivity in the brain using resting-state fMRI would allow us to probe what types of functional fluctuations in the brain can be linked to the locations of white matter bundles. By using the terminus points of white matter tracts as functional seeds, we demonstrate the ability of fMRI to identify the same connections as DTI tractography. Assessment of cortical and fMRI related changes with learning Carried out by Adam Thomas while collaborating as a student with Heidi Johansen-Berg of Oxford. In recent years, several studies have shown changes in MRI signal intensity (presumably gray matter changes) associated with specific learning tasks (i.e. juggling). We are carrying out a similar study but the focus is to understand precisely what is changing. We are observing genetic composition, myelin concentration, and blood volume. The longitudinal tasks will include exercise and juggling. We will also be pooling these data as a function of sleep-wake cycle. The hypothesis is that gray matter might vary over the course of a day. We also plan to carry this out at 7T later this year. Flip Angle Selection effects on physiologic noise and BOLD sensitivity. Carried out by Javier Gonzalez-Castillo collaborating with Jerzy Bodurka. Here we have found that changing the flip angle in fMRI time series data collection has little effect on the sensitivity across a wide range of flip angles. In the present project we investigate, both theoretically and experimentally, how TSNR dependence on flip angle varies as a function of experimental signal to noise levels (SNR) and physiological noise levels. Our results suggest that although TSNR is maximized at the Ernst angle;when physiological noise dominates, there is a wide range of angles below the Ernst angle for which TSNR suffers a negligible decrease from its maximum value. Imaging at angle below the Ernst angle might be recommended as it comes accompanied by important practical benefits such as: higher specificity of fMRI;better tissue contrast;lower susceptibility to through-plane motion;and reduced levels of heat deposition in tissue. Multi-echo fMRI development Carried out by Jennifer Evans. Most fMRI scans capture images at a single echo time that is chosen to maximize the BOLD signal. It is possible, however, to acquire several echoes in about same amount of time using a multi-echo sequence. The acquisition of more than one echo enables an estimation of T2*, the contrast responsible for the BOLD effect, while also producing a set of images with different T2* contrasts. We are investigating the use of this set of multi-echo images to separate physiological noise and other artifacts from BOLD response effects. We are also exploring a variation of this multi-echo sequence, called echo relaxation imaging (ERI) in which an entire relaxation curve is sampled in one shot. Multiple shots make up one image (64 shots per image). This will enable precise assessment of relaxation curve characteristics so that quantitative physiologic information may be extracted. Respiration noise Carried out by Jennifer Evans Physiological noise from respiration and cardiac function is present in the scans of all living subjects and can mimic activation in fMRI timeseries as well as reduce SNR and image quality in multi-shot anatomic images. Additionally, physiological noise contributions increase with higher magnetic fields, which are becoming of more interest. Both prospective and retrospective correction methods exist to attempt removal of noise effects occurring primarily from motion, cardiac function and respiration. A simulation of effects of respiration noise is being used to more systematically look at the image degradation caused by its presence and timing with respect to the imaging parameters. Informative spatial frequency in spatial pattern analysis of fMRI activation Carried out by Masaya Misaki. In multi-voxel pattern analysis, fine-scale response fMRI patterns carry significant information regarding brain function. We evaluated information of ocular dominant column responses in the early visual cortex in different spatial frequencies and found individual differences in the smoothing effect;the spatial smoothing did not affect the information for some subjects, but it reduced the information for other subjects. Disambiguating the role of human hippocampal theta oscillations Carried out by Raphael Kaplan(co-mentor Neil Burgess. In rodents, activity in the hippocampal theta frequency band has been linked to translation motion and environmental novelty. We developed a virtual navigation task with magnetoencephalography (MEG) to further investigate these effects in humans. Using MEG source reconstruction, we found that in the theta band the hippocampus is involved in the initiation of virtual movement during the task. Initial analyses comparing hippocampal activity during movement in novel and familiar environments show increased theta power in novel environments.
Showing the most recent 10 out of 84 publications
